Model Comparison
Compare Models
Select up to 3 models to compare side by side.
Qwen32B
DeepSeek32B
Llama70B
Parameters
32B
32B
70B
Minimum RAM
24GB
24GB
48GB
Recommended hardware
Mac Mini M4 Pro 48GB
Mac Mini M4 Pro 48GB
Mac Mini M4 Pro 48GB or Mac Studio
Best use cases
Chat, Reasoning, Daily use
Reasoning, Math, Analysis
Chat, Reasoning, Writing
Speed tier
Balanced
Balanced
Powerful
Comparable cloud model
Claude Sonnet / GPT-4o-mini
o1-mini
GPT-4o
~Tokens/sec (Mac Mini 48GB)
~22 tok/s
~20 tok/s
~20 tok/s
License
Apache 2.0
MIT
Meta Community License
Not sure which combination is right for your hardware? Book a consultation — we select and configure the right stack for your Mac.
We install the right combination for your hardware and workflow.
Book a consultation and we'll configure your full model stack — curated, tested, and handed back ready to use.
Book a Consultation